Public benchmarks · Raw · Reproducible

OctOpus benchmarks.

These are the current OctOpus results on public datasets where a real external reference exists. We keep the raw deltas visible against DataRobot, AutoGluon, H2O Driverless, Pecan, and Kaggle-grade baselines. Some datasets are wins, some are matches, and some are below. The page is meant to stay auditable, not flattering.

See the leaderboard

Leaderboard

Reference-backed datasets only · latest run per dataset

Dataset	Task	Winner	Score	vs Baselines	Runtime	Status
Loading…

Reproduce benchmark

If you want to reproduce these numbers, use OctOpus directly in the product. Run the same dataset, the same target, and the same objective in either the desktop app or the online version, then compare the final metric against the benchmark baseline.

# Desktop or online OctOpus
1. Open OctOpus
2. Upload the same dataset used in the benchmark
3. Use the same target column and prediction goal
4. Run the full research loop
5. Compare the final metric against the benchmark baseline